This should be easy enough to implement, but the issue is that our current implementations of Maze and ThickMaze rely on a border, so some thought will have to be given to how to handle this: ideally, we could modify these, or this is where GraphMaze could come in handy, if we can redefine our algorithms in those terms instead of in Maze terms, which should be doable.
Then GraphMaze would actually become the primary class of importance containing the maze generators, with Maze and ThickMaze being the results of homomorphisms.
This should be easy enough to implement, but the issue is that our current implementations of
Maze
andThickMaze
rely on a border, so some thought will have to be given to how to handle this: ideally, we could modify these, or this is whereGraphMaze
could come in handy, if we can redefine our algorithms in those terms instead of inMaze
terms, which should be doable.Then
GraphMaze
would actually become the primary class of importance containing the maze generators, withMaze
andThickMaze
being the results of homomorphisms.